Merge git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable

* git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable: (24 commits)
  Btrfs: fix free space cache leak
  Btrfs: avoid taking the chunk_mutex in do_chunk_alloc
  Btrfs end_bio_extent_readpage should look for locked bits
  Btrfs: don't force chunk allocation in find_free_extent
  Btrfs: Check validity before setting an acl
  Btrfs: Fix incorrect inode nlink in btrfs_link()
  Btrfs: Check if btrfs_next_leaf() returns error in btrfs_real_readdir()
  Btrfs: Check if btrfs_next_leaf() returns error in btrfs_listxattr()
  Btrfs: make uncache_state unconditional
  btrfs: using cached extent_state in set/unlock combinations
  Btrfs: avoid taking the trans_mutex in btrfs_end_transaction
  Btrfs: fix subvolume mount by name problem when default mount subvolume is set
  fix user annotation in ioctl.c
  Btrfs: check for duplicate iov_base's when doing dio reads
  btrfs: properly handle overlapping areas in memmove_extent_buffer
  Btrfs: fix memory leaks in btrfs_new_inode()
  Btrfs: check for duplicate iov_base's when doing dio reads
  Btrfs: reuse the extent_map we found when calling btrfs_get_extent
  Btrfs: do not use async submit for small DIO io's
  Btrfs: don't split dio bios if we don't have to
  ...

fs/btrfs/acl.c

@@ -178,16 +178,17 @@ static int btrfs_xattr_acl_set(struct dentry *dentry, const char *name,
 
 	if (value) {
 		acl = posix_acl_from_xattr(value, size);
-		if (acl == NULL) {
-			value = NULL;
-			size = 0;
+		if (acl) {
+			ret = posix_acl_valid(acl);
+			if (ret)
+				goto out;
 		} else if (IS_ERR(acl)) {
 			return PTR_ERR(acl);
 		}
 	}
 
 	ret = btrfs_set_acl(NULL, dentry->d_inode, acl, type);
-
+out:
 	posix_acl_release(acl);
 
 	return ret;

fs/btrfs/ctree.h

@@ -740,8 +740,10 @@ struct btrfs_space_info {
 	 */
 	unsigned long reservation_progress;
 
-	int full;		/* indicates that we cannot allocate any more
+	int full:1;		/* indicates that we cannot allocate any more
 				   chunks for this space */
+	int chunk_alloc:1;	/* set if we are allocating a chunk */
+
 	int force_alloc;	/* set if we need to force a chunk alloc for
 				   this space */
 
@@ -2576,6 +2578,11 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 			      struct inode *inode, u64 start, u64 end);
 int btrfs_release_file(struct inode *inode, struct file *file);
+void btrfs_drop_pages(struct page **pages, size_t num_pages);
+int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
+		      struct page **pages, size_t num_pages,
+		      loff_t pos, size_t write_bytes,
+		      struct extent_state **cached);
 
 /* tree-defrag.c */
 int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,

fs/btrfs/disk-io.c

@@ -3057,7 +3057,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root)
 		btrfs_destroy_pinned_extent(root,
 					    root->fs_info->pinned_extents);
 
-		t->use_count = 0;
+		atomic_set(&t->use_count, 0);
 		list_del_init(&t->list);
 		memset(t, 0, sizeof(*t));
 		kmem_cache_free(btrfs_transaction_cachep, t);

fs/btrfs/extent-tree.c

@@ -33,6 +33,25 @@
 #include "locking.h"
 #include "free-space-cache.h"
 
+/* control flags for do_chunk_alloc's force field
+ * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk
+ * if we really need one.
+ *
+ * CHUNK_ALLOC_FORCE means it must try to allocate one
+ *
+ * CHUNK_ALLOC_LIMITED means to only try and allocate one
+ * if we have very few chunks already allocated.  This is
+ * used as part of the clustering code to help make sure
+ * we have a good pool of storage to cluster in, without
+ * filling the FS with empty chunks
+ *
+ */
+enum {
+	CHUNK_ALLOC_NO_FORCE = 0,
+	CHUNK_ALLOC_FORCE = 1,
+	CHUNK_ALLOC_LIMITED = 2,
+};
+
 static int update_block_group(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      u64 bytenr, u64 num_bytes, int alloc);
@@ -3019,7 +3038,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->bytes_readonly = 0;
 	found->bytes_may_use = 0;
 	found->full = 0;
-	found->force_alloc = 0;
+	found->force_alloc = CHUNK_ALLOC_NO_FORCE;
+	found->chunk_alloc = 0;
 	*space_info = found;
 	list_add_rcu(&found->list, &info->space_info);
 	atomic_set(&found->caching_threads, 0);
@@ -3150,7 +3170,7 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
 		if (!data_sinfo->full && alloc_chunk) {
 			u64 alloc_target;
 
-			data_sinfo->force_alloc = 1;
+			data_sinfo->force_alloc = CHUNK_ALLOC_FORCE;
 			spin_unlock(&data_sinfo->lock);
 alloc:
 			alloc_target = btrfs_get_alloc_profile(root, 1);
@@ -3160,7 +3180,8 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
 
 			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
 					     bytes + 2 * 1024 * 1024,
-					     alloc_target, 0);
+					     alloc_target,
+					     CHUNK_ALLOC_NO_FORCE);
 			btrfs_end_transaction(trans, root);
 			if (ret < 0) {
 				if (ret != -ENOSPC)
@@ -3239,31 +3260,56 @@ static void force_metadata_allocation(struct btrfs_fs_info *info)
 	rcu_read_lock();
 	list_for_each_entry_rcu(found, head, list) {
 		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
-			found->force_alloc = 1;
+			found->force_alloc = CHUNK_ALLOC_FORCE;
 	}
 	rcu_read_unlock();
 }
 
 static int should_alloc_chunk(struct btrfs_root *root,
-			      struct btrfs_space_info *sinfo, u64 alloc_bytes)
+			      struct btrfs_space_info *sinfo, u64 alloc_bytes,
+			      int force)
 {
 	u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly;
+	u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved;
 	u64 thresh;
 
-	if (sinfo->bytes_used + sinfo->bytes_reserved +
-	    alloc_bytes + 256 * 1024 * 1024 < num_bytes)
+	if (force == CHUNK_ALLOC_FORCE)
+		return 1;
+
+	/*
+	 * in limited mode, we want to have some free space up to
+	 * about 1% of the FS size.
+	 */
+	if (force == CHUNK_ALLOC_LIMITED) {
+		thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+		thresh = max_t(u64, 64 * 1024 * 1024,
+			       div_factor_fine(thresh, 1));
+
+		if (num_bytes - num_allocated < thresh)
+			return 1;
+	}
+
+	/*
+	 * we have two similar checks here, one based on percentage
+	 * and once based on a hard number of 256MB.  The idea
+	 * is that if we have a good amount of free
+	 * room, don't allocate a chunk.  A good mount is
+	 * less than 80% utilized of the chunks we have allocated,
+	 * or more than 256MB free
+	 */
+	if (num_allocated + alloc_bytes + 256 * 1024 * 1024 < num_bytes)
 		return 0;
 
-	if (sinfo->bytes_used + sinfo->bytes_reserved +
-	    alloc_bytes < div_factor(num_bytes, 8))
+	if (num_allocated + alloc_bytes < div_factor(num_bytes, 8))
 		return 0;
 
 	thresh = btrfs_super_total_bytes(&root->fs_info->super_copy);
+
+	/* 256MB or 5% of the FS */
 	thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5));
 
 	if (num_bytes > thresh && sinfo->bytes_used < div_factor(num_bytes, 3))
 		return 0;
-
 	return 1;
 }
 
@@ -3273,10 +3319,9 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_space_info *space_info;
 	struct btrfs_fs_info *fs_info = extent_root->fs_info;
+	int wait_for_alloc = 0;
 	int ret = 0;
 
-	mutex_lock(&fs_info->chunk_mutex);
-
 	flags = btrfs_reduce_alloc_profile(extent_root, flags);
 
 	space_info = __find_space_info(extent_root->fs_info, flags);
@@ -3287,21 +3332,40 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 	}
 	BUG_ON(!space_info);
 
+again:
 	spin_lock(&space_info->lock);
 	if (space_info->force_alloc)
-		force = 1;
+		force = space_info->force_alloc;
 	if (space_info->full) {
 		spin_unlock(&space_info->lock);
-		goto out;
+		return 0;
 	}
 
-	if (!force && !should_alloc_chunk(extent_root, space_info,
-					  alloc_bytes)) {
+	if (!should_alloc_chunk(extent_root, space_info, alloc_bytes, force)) {
 		spin_unlock(&space_info->lock);
-		goto out;
+		return 0;
+	} else if (space_info->chunk_alloc) {
+		wait_for_alloc = 1;
+	} else {
+		space_info->chunk_alloc = 1;
 	}
+
 	spin_unlock(&space_info->lock);
 
+	mutex_lock(&fs_info->chunk_mutex);
+
+	/*
+	 * The chunk_mutex is held throughout the entirety of a chunk
+	 * allocation, so once we've acquired the chunk_mutex we know that the
+	 * other guy is done and we need to recheck and see if we should
+	 * allocate.
+	 */
+	if (wait_for_alloc) {
+		mutex_unlock(&fs_info->chunk_mutex);
+		wait_for_alloc = 0;
+		goto again;
+	}
+
 	/*
 	 * If we have mixed data/metadata chunks we want to make sure we keep
 	 * allocating mixed chunks instead of individual chunks.
@@ -3327,9 +3391,10 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 		space_info->full = 1;
 	else
 		ret = 1;
-	space_info->force_alloc = 0;
+
+	space_info->force_alloc = CHUNK_ALLOC_NO_FORCE;
+	space_info->chunk_alloc = 0;
 	spin_unlock(&space_info->lock);
-out:
 	mutex_unlock(&extent_root->fs_info->chunk_mutex);
 	return ret;
 }
@@ -5303,11 +5368,13 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
 
 		if (allowed_chunk_alloc) {
 			ret = do_chunk_alloc(trans, root, num_bytes +
-					     2 * 1024 * 1024, data, 1);
+					     2 * 1024 * 1024, data,
+					     CHUNK_ALLOC_LIMITED);
 			allowed_chunk_alloc = 0;
 			done_chunk_alloc = 1;
-		} else if (!done_chunk_alloc) {
-			space_info->force_alloc = 1;
+		} else if (!done_chunk_alloc &&
+			   space_info->force_alloc == CHUNK_ALLOC_NO_FORCE) {
+			space_info->force_alloc = CHUNK_ALLOC_LIMITED;
 		}
 
 		if (loop < LOOP_NO_EMPTY_SIZE) {
@@ -5393,7 +5460,8 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 	 */
 	if (empty_size || root->ref_cows)
 		ret = do_chunk_alloc(trans, root->fs_info->extent_root,
-				     num_bytes + 2 * 1024 * 1024, data, 0);
+				     num_bytes + 2 * 1024 * 1024, data,
+				     CHUNK_ALLOC_NO_FORCE);
 
 	WARN_ON(num_bytes < root->sectorsize);
 	ret = find_free_extent(trans, root, num_bytes, empty_size,
@@ -5405,7 +5473,7 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
 		num_bytes = num_bytes & ~(root->sectorsize - 1);
 		num_bytes = max(num_bytes, min_alloc_size);
 		do_chunk_alloc(trans, root->fs_info->extent_root,
-			       num_bytes, data, 1);
+			       num_bytes, data, CHUNK_ALLOC_FORCE);
 		goto again;
 	}
 	if (ret == -ENOSPC && btrfs_test_opt(root, ENOSPC_DEBUG)) {
@@ -8109,13 +8177,15 @@ int btrfs_set_block_group_ro(struct btrfs_root *root,
 
 	alloc_flags = update_block_group_flags(root, cache->flags);
 	if (alloc_flags != cache->flags)
-		do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1);
+		do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
+			       CHUNK_ALLOC_FORCE);
 
 	ret = set_block_group_ro(cache);
 	if (!ret)
 		goto out;
 	alloc_flags = get_alloc_profile(root, cache->space_info->flags);
-	ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1);
+	ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
+			     CHUNK_ALLOC_FORCE);
 	if (ret < 0)
 		goto out;
 	ret = set_block_group_ro(cache);
@@ -8128,7 +8198,8 @@ int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans,
 			    struct btrfs_root *root, u64 type)
 {
 	u64 alloc_flags = get_alloc_profile(root, type);
-	return do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, 1);
+	return do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags,
+			      CHUNK_ALLOC_FORCE);
 }
 
 /*

fs/btrfs/extent_io.c

@@ -690,6 +690,15 @@ static void cache_state(struct extent_state *state,
 	}
 }
 
+static void uncache_state(struct extent_state **cached_ptr)
+{
+	if (cached_ptr && (*cached_ptr)) {
+		struct extent_state *state = *cached_ptr;
+		*cached_ptr = NULL;
+		free_extent_state(state);
+	}
+}
+
 /*
  * set some bits on a range in the tree.  This may require allocations or
  * sleeping, so the gfp mask is used to indicate what is allowed.
@@ -940,10 +949,10 @@ static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 }
 
 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
-			gfp_t mask)
+			struct extent_state **cached_state, gfp_t mask)
 {
-	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
-			      NULL, mask);
+	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0,
+			      NULL, cached_state, mask);
 }
 
 static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
@@ -1012,8 +1021,7 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
 				mask);
 }
 
-int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end,
-		  gfp_t mask)
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
 {
 	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL,
 				mask);
@@ -1735,6 +1743,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 
 	do {
 		struct page *page = bvec->bv_page;
+		struct extent_state *cached = NULL;
+		struct extent_state *state;
+
 		tree = &BTRFS_I(page->mapping->host)->io_tree;
 
 		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
@@ -1749,9 +1760,20 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 		if (++bvec <= bvec_end)
 			prefetchw(&bvec->bv_page->flags);
 
+		spin_lock(&tree->lock);
+		state = find_first_extent_bit_state(tree, start, EXTENT_LOCKED);
+		if (state && state->start == start) {
+			/*
+			 * take a reference on the state, unlock will drop
+			 * the ref
+			 */
+			cache_state(state, &cached);
+		}
+		spin_unlock(&tree->lock);
+
 		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
 			ret = tree->ops->readpage_end_io_hook(page, start, end,
-							      NULL);
+							      state);
 			if (ret)
 				uptodate = 0;
 		}
@@ -1764,15 +1786,16 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 					test_bit(BIO_UPTODATE, &bio->bi_flags);
 				if (err)
 					uptodate = 0;
+				uncache_state(&cached);
 				continue;
 			}
 		}
 
 		if (uptodate) {
-			set_extent_uptodate(tree, start, end,
+			set_extent_uptodate(tree, start, end, &cached,
 					    GFP_ATOMIC);
 		}
-		unlock_extent(tree, start, end, GFP_ATOMIC);
+		unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
 
 		if (whole_page) {
 			if (uptodate) {
@@ -1811,6 +1834,7 @@ static void end_bio_extent_preparewrite(struct bio *bio, int err)
 
 	do {
 		struct page *page = bvec->bv_page;
+		struct extent_state *cached = NULL;
 		tree = &BTRFS_I(page->mapping->host)->io_tree;
 
 		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
@@ -1821,13 +1845,14 @@ static void end_bio_extent_preparewrite(struct bio *bio, int err)
 			prefetchw(&bvec->bv_page->flags);
 
 		if (uptodate) {
-			set_extent_uptodate(tree, start, end, GFP_ATOMIC);
+			set_extent_uptodate(tree, start, end, &cached,
+					    GFP_ATOMIC);
 		} else {
 			ClearPageUptodate(page);
 			SetPageError(page);
 		}
 
-		unlock_extent(tree, start, end, GFP_ATOMIC);
+		unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
 
 	} while (bvec >= bio->bi_io_vec);
 
@@ -2016,14 +2041,17 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	while (cur <= end) {
 		if (cur >= last_byte) {
 			char *userpage;
+			struct extent_state *cached = NULL;
+
 			iosize = PAGE_CACHE_SIZE - page_offset;
 			userpage = kmap_atomic(page, KM_USER0);
 			memset(userpage + page_offset, 0, iosize);
 			flush_dcache_page(page);
 			kunmap_atomic(userpage, KM_USER0);
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
-					    GFP_NOFS);
-			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+					    &cached, GFP_NOFS);
+			unlock_extent_cached(tree, cur, cur + iosize - 1,
+					     &cached, GFP_NOFS);
 			break;
 		}
 		em = get_extent(inode, page, page_offset, cur,
@@ -2063,14 +2091,17 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		/* we've found a hole, just zero and go on */
 		if (block_start == EXTENT_MAP_HOLE) {
 			char *userpage;
+			struct extent_state *cached = NULL;
+
 			userpage = kmap_atomic(page, KM_USER0);
 			memset(userpage + page_offset, 0, iosize);
 			flush_dcache_page(page);
 			kunmap_atomic(userpage, KM_USER0);
 
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
-					    GFP_NOFS);
-			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+					    &cached, GFP_NOFS);
+			unlock_extent_cached(tree, cur, cur + iosize - 1,
+			                     &cached, GFP_NOFS);
 			cur = cur + iosize;
 			page_offset += iosize;
 			continue;
@@ -2789,9 +2820,12 @@ int extent_prepare_write(struct extent_io_tree *tree,
 			iocount++;
 			block_start = block_start + iosize;
 		} else {
-			set_extent_uptodate(tree, block_start, cur_end,
+			struct extent_state *cached = NULL;
+
+			set_extent_uptodate(tree, block_start, cur_end, &cached,
 					    GFP_NOFS);
-			unlock_extent(tree, block_start, cur_end, GFP_NOFS);
+			unlock_extent_cached(tree, block_start, cur_end,
+					     &cached, GFP_NOFS);
 			block_start = cur_end + 1;
 		}
 		page_offset = block_start & (PAGE_CACHE_SIZE - 1);
@@ -3457,7 +3491,7 @@ int set_extent_buffer_uptodate(struct extent_io_tree *tree,
 	num_pages = num_extent_pages(eb->start, eb->len);
 
 	set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
-			    GFP_NOFS);
+			    NULL, GFP_NOFS);
 	for (i = 0; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
 		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
@@ -3885,6 +3919,12 @@ static void move_pages(struct page *dst_page, struct page *src_page,
 	kunmap_atomic(dst_kaddr, KM_USER0);
 }
 
+static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len)
+{
+	unsigned long distance = (src > dst) ? src - dst : dst - src;
+	return distance < len;
+}
+
 static void copy_pages(struct page *dst_page, struct page *src_page,
 		       unsigned long dst_off, unsigned long src_off,
 		       unsigned long len)
@@ -3892,10 +3932,12 @@ static void copy_pages(struct page *dst_page, struct page *src_page,
 	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
 	char *src_kaddr;
 
-	if (dst_page != src_page)
+	if (dst_page != src_page) {
 		src_kaddr = kmap_atomic(src_page, KM_USER1);
-	else
+	} else {
 		src_kaddr = dst_kaddr;
+		BUG_ON(areas_overlap(src_off, dst_off, len));
+	}
 
 	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
 	kunmap_atomic(dst_kaddr, KM_USER0);
@@ -3970,7 +4012,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 		       "len %lu len %lu\n", dst_offset, len, dst->len);
 		BUG_ON(1);
 	}
-	if (dst_offset < src_offset) {
+	if (!areas_overlap(src_offset, dst_offset, len)) {
 		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
 		return;
 	}

fs/btrfs/extent_io.h

@@ -208,7 +208,7 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		   int bits, int exclusive_bits, u64 *failed_start,
 		   struct extent_state **cached_state, gfp_t mask);
 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
-			gfp_t mask);
+			struct extent_state **cached_state, gfp_t mask);
 int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 		   gfp_t mask);
 int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,

fs/btrfs/file.c

@@ -104,7 +104,7 @@ static noinline int btrfs_copy_from_user(loff_t pos, int num_pages,
 /*
  * unlocks pages after btrfs_file_write is done with them
  */
-static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages)
+void btrfs_drop_pages(struct page **pages, size_t num_pages)
 {
 	size_t i;
 	for (i = 0; i < num_pages; i++) {
@@ -127,16 +127,13 @@ static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages)
  * this also makes the decision about creating an inline extent vs
  * doing real data extents, marking pages dirty and delalloc as required.
  */
-static noinline int dirty_and_release_pages(struct btrfs_root *root,
-					    struct file *file,
-					    struct page **pages,
-					    size_t num_pages,
-					    loff_t pos,
-					    size_t write_bytes)
+int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
+		      struct page **pages, size_t num_pages,
+		      loff_t pos, size_t write_bytes,
+		      struct extent_state **cached)
 {
 	int err = 0;
 	int i;
-	struct inode *inode = fdentry(file)->d_inode;
 	u64 num_bytes;
 	u64 start_pos;
 	u64 end_of_last_block;
@@ -149,7 +146,7 @@ static noinline int dirty_and_release_pages(struct btrfs_root *root,
 
 	end_of_last_block = start_pos + num_bytes - 1;
 	err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
-					NULL);
+					cached);
 	if (err)
 		return err;
 
@@ -992,9 +989,9 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 		}
 
 		if (copied > 0) {
-			ret = dirty_and_release_pages(root, file, pages,
-						      dirty_pages, pos,
-						      copied);
+			ret = btrfs_dirty_pages(root, inode, pages,
+						dirty_pages, pos, copied,
+						NULL);
 			if (ret) {
 				btrfs_delalloc_release_space(inode,
 					dirty_pages << PAGE_CACHE_SHIFT);

fs/btrfs/free-space-cache.c

@@ -508,6 +508,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	struct inode *inode;
 	struct rb_node *node;
 	struct list_head *pos, *n;
+	struct page **pages;
 	struct page *page;
 	struct extent_state *cached_state = NULL;
 	struct btrfs_free_cluster *cluster = NULL;
@@ -517,13 +518,13 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	u64 start, end, len;
 	u64 bytes = 0;
 	u32 *crc, *checksums;
-	pgoff_t index = 0, last_index = 0;
 	unsigned long first_page_offset;
-	int num_checksums;
+	int index = 0, num_pages = 0;
 	int entries = 0;
 	int bitmaps = 0;
 	int ret = 0;
 	bool next_page = false;
+	bool out_of_space = false;
 
 	root = root->fs_info->tree_root;
 
@@ -551,24 +552,31 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		return 0;
 	}
 
-	last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
+	num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
+		PAGE_CACHE_SHIFT;
 	filemap_write_and_wait(inode->i_mapping);
 	btrfs_wait_ordered_range(inode, inode->i_size &
 				 ~(root->sectorsize - 1), (u64)-1);
 
 	/* We need a checksum per page. */
-	num_checksums = i_size_read(inode) / PAGE_CACHE_SIZE;
-	crc = checksums  = kzalloc(sizeof(u32) * num_checksums, GFP_NOFS);
+	crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS);
 	if (!crc) {
 		iput(inode);
 		return 0;
 	}
 
+	pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS);
+	if (!pages) {
+		kfree(crc);
+		iput(inode);
+		return 0;
+	}
+
 	/* Since the first page has all of our checksums and our generation we
 	 * need to calculate the offset into the page that we can start writing
 	 * our entries.
 	 */
-	first_page_offset = (sizeof(u32) * num_checksums) + sizeof(u64);
+	first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64);
 
 	/* Get the cluster for this block_group if it exists */
 	if (!list_empty(&block_group->cluster_list))
@@ -590,20 +598,18 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 	 * after find_get_page at this point.  Just putting this here so people
 	 * know and don't freak out.
 	 */
-	while (index <= last_index) {
+	while (index < num_pages) {
 		page = grab_cache_page(inode->i_mapping, index);
 		if (!page) {
-			pgoff_t i = 0;
+			int i;
 
-			while (i < index) {
-				page = find_get_page(inode->i_mapping, i);
-				unlock_page(page);
-				page_cache_release(page);
-				page_cache_release(page);
-				i++;
+			for (i = 0; i < num_pages; i++) {
+				unlock_page(pages[i]);
+				page_cache_release(pages[i]);
 			}
 			goto out_free;
 		}
+		pages[index] = page;
 		index++;
 	}
 
@@ -631,7 +637,12 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 			offset = start_offset;
 		}
 
-		page = find_get_page(inode->i_mapping, index);
+		if (index >= num_pages) {
+			out_of_space = true;
+			break;
+		}
+
+		page = pages[index];
 
 		addr = kmap(page);
 		entry = addr + start_offset;
@@ -708,23 +719,6 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 
 		bytes += PAGE_CACHE_SIZE;
 
-		ClearPageChecked(page);
-		set_page_extent_mapped(page);
-		SetPageUptodate(page);
-		set_page_dirty(page);
-
-		/*
-		 * We need to release our reference we got for grab_cache_page,
-		 * except for the first page which will hold our checksums, we
-		 * do that below.
-		 */
-		if (index != 0) {
-			unlock_page(page);
-			page_cache_release(page);
-		}
-
-		page_cache_release(page);
-
 		index++;
 	} while (node || next_page);
 
@@ -734,7 +728,11 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		struct btrfs_free_space *entry =
 			list_entry(pos, struct btrfs_free_space, list);
 
-		page = find_get_page(inode->i_mapping, index);
+		if (index >= num_pages) {
+			out_of_space = true;
+			break;
+		}
+		page = pages[index];
 
 		addr = kmap(page);
 		memcpy(addr, entry->bitmap, PAGE_CACHE_SIZE);
@@ -745,64 +743,58 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		crc++;
 		bytes += PAGE_CACHE_SIZE;
 
-		ClearPageChecked(page);
-		set_page_extent_mapped(page);
-		SetPageUptodate(page);
-		set_page_dirty(page);
-		unlock_page(page);
-		page_cache_release(page);
-		page_cache_release(page);
 		list_del_init(&entry->list);
 		index++;
 	}
 
+	if (out_of_space) {
+		btrfs_drop_pages(pages, num_pages);
+		unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
+				     i_size_read(inode) - 1, &cached_state,
+				     GFP_NOFS);
+		ret = 0;
+		goto out_free;
+	}
+
 	/* Zero out the rest of the pages just to make sure */
-	while (index <= last_index) {
+	while (index < num_pages) {
 		void *addr;
 
-		page = find_get_page(inode->i_mapping, index);
-
+		page = pages[index];
 		addr = kmap(page);
 		memset(addr, 0, PAGE_CACHE_SIZE);
 		kunmap(page);
-		ClearPageChecked(page);
-		set_page_extent_mapped(page);
-		SetPageUptodate(page);
-		set_page_dirty(page);
-		unlock_page(page);
-		page_cache_release(page);
-		page_cache_release(page);
 		bytes += PAGE_CACHE_SIZE;
 		index++;
 	}
 
-	btrfs_set_extent_delalloc(inode, 0, bytes - 1, &cached_state);
-
 	/* Write the checksums and trans id to the first page */
 	{
 		void *addr;
 		u64 *gen;
 
-		page = find_get_page(inode->i_mapping, 0);
+		page = pages[0];
 
 		addr = kmap(page);
-		memcpy(addr, checksums, sizeof(u32) * num_checksums);
-		gen = addr + (sizeof(u32) * num_checksums);
+		memcpy(addr, checksums, sizeof(u32) * num_pages);
+		gen = addr + (sizeof(u32) * num_pages);
 		*gen = trans->transid;
 		kunmap(page);
-		ClearPageChecked(page);
-		set_page_extent_mapped(page);
-		SetPageUptodate(page);
-		set_page_dirty(page);
-		unlock_page(page);
-		page_cache_release(page);
-		page_cache_release(page);
 	}
-	BTRFS_I(inode)->generation = trans->transid;
 
+	ret = btrfs_dirty_pages(root, inode, pages, num_pages, 0,
+					    bytes, &cached_state);
+	btrfs_drop_pages(pages, num_pages);
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
 			     i_size_read(inode) - 1, &cached_state, GFP_NOFS);
 
+	if (ret) {
+		ret = 0;
+		goto out_free;
+	}
+
+	BTRFS_I(inode)->generation = trans->transid;
+
 	filemap_write_and_wait(inode->i_mapping);
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
@@ -853,6 +845,7 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 		BTRFS_I(inode)->generation = 0;
 	}
 	kfree(checksums);
+	kfree(pages);
 	btrfs_update_inode(trans, root, inode);
 	iput(inode);
 	return ret;

fs/btrfs/inode.c

@@ -1770,9 +1770,12 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 	add_pending_csums(trans, inode, ordered_extent->file_offset,
 			  &ordered_extent->list);
 
-	btrfs_ordered_update_i_size(inode, 0, ordered_extent);
-	ret = btrfs_update_inode(trans, root, inode);
-	BUG_ON(ret);
+	ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
+	if (!ret) {
+		ret = btrfs_update_inode(trans, root, inode);
+		BUG_ON(ret);
+	}
+	ret = 0;
 out:
 	if (nolock) {
 		if (trans)
@@ -2590,6 +2593,13 @@ static void fill_inode_item(struct btrfs_trans_handle *trans,
 			    struct btrfs_inode_item *item,
 			    struct inode *inode)
 {
+	if (!leaf->map_token)
+		map_private_extent_buffer(leaf, (unsigned long)item,
+					  sizeof(struct btrfs_inode_item),
+					  &leaf->map_token, &leaf->kaddr,
+					  &leaf->map_start, &leaf->map_len,
+					  KM_USER1);
+
 	btrfs_set_inode_uid(leaf, item, inode->i_uid);
 	btrfs_set_inode_gid(leaf, item, inode->i_gid);
 	btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
@@ -2618,6 +2628,11 @@ static void fill_inode_item(struct btrfs_trans_handle *trans,
 	btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
 	btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
 	btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group);
+
+	if (leaf->map_token) {
+		unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+		leaf->map_token = NULL;
+	}
 }
 
 /*
@@ -4207,10 +4222,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 	struct btrfs_key found_key;
 	struct btrfs_path *path;
 	int ret;
-	u32 nritems;
 	struct extent_buffer *leaf;
 	int slot;
-	int advance;
 	unsigned char d_type;
 	int over = 0;
 	u32 di_cur;
@@ -4253,27 +4266,19 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
-	advance = 0;
 
 	while (1) {
 		leaf = path->nodes[0];
-		nritems = btrfs_header_nritems(leaf);
 		slot = path->slots[0];
-		if (advance || slot >= nritems) {
-			if (slot >= nritems - 1) {
-				ret = btrfs_next_leaf(root, path);
-				if (ret)
-					break;
-				leaf = path->nodes[0];
-				nritems = btrfs_header_nritems(leaf);
-				slot = path->slots[0];
-			} else {
-				slot++;
-				path->slots[0]++;
-			}
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto err;
+			else if (ret > 0)
+				break;
+			continue;
 		}
 
-		advance = 1;
 		item = btrfs_item_nr(leaf, slot);
 		btrfs_item_key_to_cpu(leaf, &found_key, slot);
 
@@ -4282,7 +4287,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 		if (btrfs_key_type(&found_key) != key_type)
 			break;
 		if (found_key.offset < filp->f_pos)
-			continue;
+			goto next;
 
 		filp->f_pos = found_key.offset;
 
@@ -4335,6 +4340,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent,
 			di_cur += di_len;
 			di = (struct btrfs_dir_item *)((char *)di + di_len);
 		}
+next:
+		path->slots[0]++;
 	}
 
 	/* Reached end of directory/root. Bump pos past the last item. */
@@ -4527,14 +4534,17 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
 	BUG_ON(!path);
 
 	inode = new_inode(root->fs_info->sb);
-	if (!inode)
+	if (!inode) {
+		btrfs_free_path(path);
 		return ERR_PTR(-ENOMEM);
+	}
 
 	if (dir) {
 		trace_btrfs_inode_request(dir);
 
 		ret = btrfs_set_inode_index(dir, index);
 		if (ret) {
+			btrfs_free_path(path);
 			iput(inode);
 			return ERR_PTR(ret);
 		}
@@ -4834,9 +4844,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	if (inode->i_nlink == ~0U)
 		return -EMLINK;
 
-	btrfs_inc_nlink(inode);
-	inode->i_ctime = CURRENT_TIME;
-
 	err = btrfs_set_inode_index(dir, &index);
 	if (err)
 		goto fail;
@@ -4852,6 +4859,9 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 		goto fail;
 	}
 
+	btrfs_inc_nlink(inode);
+	inode->i_ctime = CURRENT_TIME;
+
 	btrfs_set_trans_block_group(trans, dir);
 	ihold(inode);
 
@@ -5221,7 +5231,7 @@ struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
 			btrfs_mark_buffer_dirty(leaf);
 		}
 		set_extent_uptodate(io_tree, em->start,
-				    extent_map_end(em) - 1, GFP_NOFS);
+				    extent_map_end(em) - 1, NULL, GFP_NOFS);
 		goto insert;
 	} else {
 		printk(KERN_ERR "btrfs unknown found_type %d\n", found_type);
@@ -5428,17 +5438,30 @@ struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *pag
 }
 
 static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
+						  struct extent_map *em,
 						  u64 start, u64 len)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_trans_handle *trans;
-	struct extent_map *em;
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
 	struct btrfs_key ins;
 	u64 alloc_hint;
 	int ret;
+	bool insert = false;
 
-	btrfs_drop_extent_cache(inode, start, start + len - 1, 0);
+	/*
+	 * Ok if the extent map we looked up is a hole and is for the exact
+	 * range we want, there is no reason to allocate a new one, however if
+	 * it is not right then we need to free this one and drop the cache for
+	 * our range.
+	 */
+	if (em->block_start != EXTENT_MAP_HOLE || em->start != start ||
+	    em->len != len) {
+		free_extent_map(em);
+		em = NULL;
+		insert = true;
+		btrfs_drop_extent_cache(inode, start, start + len - 1, 0);
+	}
 
 	trans = btrfs_join_transaction(root, 0);
 	if (IS_ERR(trans))
@@ -5454,10 +5477,12 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
 		goto out;
 	}
 
-	em = alloc_extent_map(GFP_NOFS);
 	if (!em) {
-		em = ERR_PTR(-ENOMEM);
-		goto out;
+		em = alloc_extent_map(GFP_NOFS);
+		if (!em) {
+			em = ERR_PTR(-ENOMEM);
+			goto out;
+		}
 	}
 
 	em->start = start;
@@ -5467,9 +5492,15 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode,
 	em->block_start = ins.objectid;
 	em->block_len = ins.offset;
 	em->bdev = root->fs_info->fs_devices->latest_bdev;
+
+	/*
+	 * We need to do this because if we're using the original em we searched
+	 * for, we could have EXTENT_FLAG_VACANCY set, and we don't want that.
+	 */
+	em->flags = 0;
 	set_bit(EXTENT_FLAG_PINNED, &em->flags);
 
-	while (1) {
+	while (insert) {
 		write_lock(&em_tree->lock);
 		ret = add_extent_mapping(em_tree, em);
 		write_unlock(&em_tree->lock);
@@ -5687,8 +5718,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	 * it above
 	 */
 	len = bh_result->b_size;
-	free_extent_map(em);
-	em = btrfs_new_extent_direct(inode, start, len);
+	em = btrfs_new_extent_direct(inode, em, start, len);
 	if (IS_ERR(em))
 		return PTR_ERR(em);
 	len = min(len, em->len - (start - em->start));
@@ -5851,8 +5881,10 @@ static void btrfs_endio_direct_write(struct bio *bio, int err)
 	}
 
 	add_pending_csums(trans, inode, ordered->file_offset, &ordered->list);
-	btrfs_ordered_update_i_size(inode, 0, ordered);
-	btrfs_update_inode(trans, root, inode);
+	ret = btrfs_ordered_update_i_size(inode, 0, ordered);
+	if (!ret)
+		btrfs_update_inode(trans, root, inode);
+	ret = 0;
 out_unlock:
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset,
 			     ordered->file_offset + ordered->len - 1,
@@ -5938,7 +5970,7 @@ static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev,
 
 static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
 					 int rw, u64 file_offset, int skip_sum,
-					 u32 *csums)
+					 u32 *csums, int async_submit)
 {
 	int write = rw & REQ_WRITE;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -5949,13 +5981,24 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
 	if (ret)
 		goto err;
 
-	if (write && !skip_sum) {
+	if (skip_sum)
+		goto map;
+
+	if (write && async_submit) {
 		ret = btrfs_wq_submit_bio(root->fs_info,
 				   inode, rw, bio, 0, 0,
 				   file_offset,
 				   __btrfs_submit_bio_start_direct_io,
 				   __btrfs_submit_bio_done);
 		goto err;
+	} else if (write) {
+		/*
+		 * If we aren't doing async submit, calculate the csum of the
+		 * bio now.
+		 */
+		ret = btrfs_csum_one_bio(root, inode, bio, file_offset, 1);
+		if (ret)
+			goto err;
 	} else if (!skip_sum) {
 		ret = btrfs_lookup_bio_sums_dio(root, inode, bio,
 					  file_offset, csums);
@@ -5963,7 +6006,8 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
 			goto err;
 	}
 
-	ret = btrfs_map_bio(root, rw, bio, 0, 1);
+map:
+	ret = btrfs_map_bio(root, rw, bio, 0, async_submit);
 err:
 	bio_put(bio);
 	return ret;
@@ -5985,15 +6029,9 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip,
 	int nr_pages = 0;
 	u32 *csums = dip->csums;
 	int ret = 0;
+	int async_submit = 0;
 	int write = rw & REQ_WRITE;
 
-	bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS);
-	if (!bio)
-		return -ENOMEM;
-	bio->bi_private = dip;
-	bio->bi_end_io = btrfs_end_dio_bio;
-	atomic_inc(&dip->pending_bios);
-
 	map_length = orig_bio->bi_size;
 	ret = btrfs_map_block(map_tree, READ, start_sector << 9,
 			      &map_length, NULL, 0);
@@ -6002,6 +6040,19 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip,
 		return -EIO;
 	}
 
+	if (map_length >= orig_bio->bi_size) {
+		bio = orig_bio;
+		goto submit;
+	}
+
+	async_submit = 1;
+	bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS);
+	if (!bio)
+		return -ENOMEM;
+	bio->bi_private = dip;
+	bio->bi_end_io = btrfs_end_dio_bio;
+	atomic_inc(&dip->pending_bios);
+
 	while (bvec <= (orig_bio->bi_io_vec + orig_bio->bi_vcnt - 1)) {
 		if (unlikely(map_length < submit_len + bvec->bv_len ||
 		    bio_add_page(bio, bvec->bv_page, bvec->bv_len,
@@ -6015,7 +6066,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip,
 			atomic_inc(&dip->pending_bios);
 			ret = __btrfs_submit_dio_bio(bio, inode, rw,
 						     file_offset, skip_sum,
-						     csums);
+						     csums, async_submit);
 			if (ret) {
 				bio_put(bio);
 				atomic_dec(&dip->pending_bios);
@@ -6052,8 +6103,9 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip,
 		}
 	}
 
+submit:
 	ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum,
-				     csums);
+				     csums, async_submit);
 	if (!ret)
 		return 0;
 
@@ -6148,6 +6200,7 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io
 			unsigned long nr_segs)
 {
 	int seg;
+	int i;
 	size_t size;
 	unsigned long addr;
 	unsigned blocksize_mask = root->sectorsize - 1;
@@ -6162,8 +6215,22 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io
 		addr = (unsigned long)iov[seg].iov_base;
 		size = iov[seg].iov_len;
 		end += size;
-		if ((addr & blocksize_mask) || (size & blocksize_mask)) 
+		if ((addr & blocksize_mask) || (size & blocksize_mask))
 			goto out;
+
+		/* If this is a write we don't need to check anymore */
+		if (rw & WRITE)
+			continue;
+
+		/*
+		 * Check to make sure we don't have duplicate iov_base's in this
+		 * iovec, if so return EINVAL, otherwise we'll get csum errors
+		 * when reading back.
+		 */
+		for (i = seg + 1; i < nr_segs; i++) {
+			if (iov[seg].iov_base == iov[i].iov_base)
+				goto out;
+		}
 	}
 	retval = 0;
 out:

fs/btrfs/ioctl.c

@@ -2287,7 +2287,7 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg)
 	struct btrfs_ioctl_space_info space;
 	struct btrfs_ioctl_space_info *dest;
 	struct btrfs_ioctl_space_info *dest_orig;
-	struct btrfs_ioctl_space_info *user_dest;
+	struct btrfs_ioctl_space_info __user *user_dest;
 	struct btrfs_space_info *info;
 	u64 types[] = {BTRFS_BLOCK_GROUP_DATA,
 		       BTRFS_BLOCK_GROUP_SYSTEM,

fs/btrfs/super.c

@@ -159,7 +159,7 @@ enum {
 	Opt_compress_type, Opt_compress_force, Opt_compress_force_type,
 	Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
 	Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
-	Opt_enospc_debug, Opt_err,
+	Opt_enospc_debug, Opt_subvolrootid, Opt_err,
 };
 
 static match_table_t tokens = {
@@ -189,6 +189,7 @@ static match_table_t tokens = {
 	{Opt_clear_cache, "clear_cache"},
 	{Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"},
 	{Opt_enospc_debug, "enospc_debug"},
+	{Opt_subvolrootid, "subvolrootid=%d"},
 	{Opt_err, NULL},
 };
 
@@ -232,6 +233,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			break;
 		case Opt_subvol:
 		case Opt_subvolid:
+		case Opt_subvolrootid:
 		case Opt_device:
 			/*
 			 * These are parsed by btrfs_parse_early_options
@@ -388,7 +390,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
  */
 static int btrfs_parse_early_options(const char *options, fmode_t flags,
 		void *holder, char **subvol_name, u64 *subvol_objectid,
-		struct btrfs_fs_devices **fs_devices)
+		u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices)
 {
 	substring_t args[MAX_OPT_ARGS];
 	char *opts, *orig, *p;
@@ -429,6 +431,18 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags,
 					*subvol_objectid = intarg;
 			}
 			break;
+		case Opt_subvolrootid:
+			intarg = 0;
+			error = match_int(&args[0], &intarg);
+			if (!error) {
+				/* we want the original fs_tree */
+				if (!intarg)
+					*subvol_rootid =
+						BTRFS_FS_TREE_OBJECTID;
+				else
+					*subvol_rootid = intarg;
+			}
+			break;
 		case Opt_device:
 			error = btrfs_scan_one_device(match_strdup(&args[0]),
 					flags, holder, fs_devices);
@@ -736,6 +750,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 	fmode_t mode = FMODE_READ;
 	char *subvol_name = NULL;
 	u64 subvol_objectid = 0;
+	u64 subvol_rootid = 0;
 	int error = 0;
 
 	if (!(flags & MS_RDONLY))
@@ -743,7 +758,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 
 	error = btrfs_parse_early_options(data, mode, fs_type,
 					  &subvol_name, &subvol_objectid,
-					  &fs_devices);
+					  &subvol_rootid, &fs_devices);
 	if (error)
 		return ERR_PTR(error);
 
@@ -807,15 +822,17 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 		s->s_flags |= MS_ACTIVE;
 	}
 
-	root = get_default_root(s, subvol_objectid);
-	if (IS_ERR(root)) {
-		error = PTR_ERR(root);
-		deactivate_locked_super(s);
-		goto error_free_subvol_name;
-	}
 	/* if they gave us a subvolume name bind mount into that */
 	if (strcmp(subvol_name, ".")) {
 		struct dentry *new_root;
+
+		root = get_default_root(s, subvol_rootid);
+		if (IS_ERR(root)) {
+			error = PTR_ERR(root);
+			deactivate_locked_super(s);
+			goto error_free_subvol_name;
+		}
+
 		mutex_lock(&root->d_inode->i_mutex);
 		new_root = lookup_one_len(subvol_name, root,
 				      strlen(subvol_name));
@@ -836,6 +853,13 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 		}
 		dput(root);
 		root = new_root;
+	} else {
+		root = get_default_root(s, subvol_objectid);
+		if (IS_ERR(root)) {
+			error = PTR_ERR(root);
+			deactivate_locked_super(s);
+			goto error_free_subvol_name;
+		}
 	}
 
 	kfree(subvol_name);

fs/btrfs/transaction.c

@@ -32,10 +32,8 @@
 
 static noinline void put_transaction(struct btrfs_transaction *transaction)
 {
-	WARN_ON(transaction->use_count == 0);
-	transaction->use_count--;
-	if (transaction->use_count == 0) {
-		list_del_init(&transaction->list);
+	WARN_ON(atomic_read(&transaction->use_count) == 0);
+	if (atomic_dec_and_test(&transaction->use_count)) {
 		memset(transaction, 0, sizeof(*transaction));
 		kmem_cache_free(btrfs_transaction_cachep, transaction);
 	}
@@ -60,14 +58,14 @@ static noinline int join_transaction(struct btrfs_root *root)
 		if (!cur_trans)
 			return -ENOMEM;
 		root->fs_info->generation++;
-		cur_trans->num_writers = 1;
+		atomic_set(&cur_trans->num_writers, 1);
 		cur_trans->num_joined = 0;
 		cur_trans->transid = root->fs_info->generation;
 		init_waitqueue_head(&cur_trans->writer_wait);
 		init_waitqueue_head(&cur_trans->commit_wait);
 		cur_trans->in_commit = 0;
 		cur_trans->blocked = 0;
-		cur_trans->use_count = 1;
+		atomic_set(&cur_trans->use_count, 1);
 		cur_trans->commit_done = 0;
 		cur_trans->start_time = get_seconds();
 
@@ -88,7 +86,7 @@ static noinline int join_transaction(struct btrfs_root *root)
 		root->fs_info->running_transaction = cur_trans;
 		spin_unlock(&root->fs_info->new_trans_lock);
 	} else {
-		cur_trans->num_writers++;
+		atomic_inc(&cur_trans->num_writers);
 		cur_trans->num_joined++;
 	}
 
@@ -145,7 +143,7 @@ static void wait_current_trans(struct btrfs_root *root)
 	cur_trans = root->fs_info->running_transaction;
 	if (cur_trans && cur_trans->blocked) {
 		DEFINE_WAIT(wait);
-		cur_trans->use_count++;
+		atomic_inc(&cur_trans->use_count);
 		while (1) {
 			prepare_to_wait(&root->fs_info->transaction_wait, &wait,
 					TASK_UNINTERRUPTIBLE);
@@ -181,6 +179,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 {
 	struct btrfs_trans_handle *h;
 	struct btrfs_transaction *cur_trans;
+	int retries = 0;
 	int ret;
 
 	if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
@@ -204,7 +203,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 	}
 
 	cur_trans = root->fs_info->running_transaction;
-	cur_trans->use_count++;
+	atomic_inc(&cur_trans->use_count);
 	if (type != TRANS_JOIN_NOLOCK)
 		mutex_unlock(&root->fs_info->trans_mutex);
 
@@ -224,10 +223,18 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
 
 	if (num_items > 0) {
 		ret = btrfs_trans_reserve_metadata(h, root, num_items);
-		if (ret == -EAGAIN) {
+		if (ret == -EAGAIN && !retries) {
+			retries++;
 			btrfs_commit_transaction(h, root);
 			goto again;
+		} else if (ret == -EAGAIN) {
+			/*
+			 * We have already retried and got EAGAIN, so really we
+			 * don't have space, so set ret to -ENOSPC.
+			 */
+			ret = -ENOSPC;
 		}
+
 		if (ret < 0) {
 			btrfs_end_transaction(h, root);
 			return ERR_PTR(ret);
@@ -327,7 +334,7 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid)
 			goto out_unlock;  /* nothing committing|committed */
 	}
 
-	cur_trans->use_count++;
+	atomic_inc(&cur_trans->use_count);
 	mutex_unlock(&root->fs_info->trans_mutex);
 
 	wait_for_commit(root, cur_trans);
@@ -457,18 +464,14 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			wake_up_process(info->transaction_kthread);
 	}
 
-	if (lock)
-		mutex_lock(&info->trans_mutex);
 	WARN_ON(cur_trans != info->running_transaction);
-	WARN_ON(cur_trans->num_writers < 1);
-	cur_trans->num_writers--;
+	WARN_ON(atomic_read(&cur_trans->num_writers) < 1);
+	atomic_dec(&cur_trans->num_writers);
 
 	smp_mb();
 	if (waitqueue_active(&cur_trans->writer_wait))
 		wake_up(&cur_trans->writer_wait);
 	put_transaction(cur_trans);
-	if (lock)
-		mutex_unlock(&info->trans_mutex);
 
 	if (current->journal_info == trans)
 		current->journal_info = NULL;
@@ -1178,7 +1181,7 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
 	/* take transaction reference */
 	mutex_lock(&root->fs_info->trans_mutex);
 	cur_trans = trans->transaction;
-	cur_trans->use_count++;
+	atomic_inc(&cur_trans->use_count);
 	mutex_unlock(&root->fs_info->trans_mutex);
 
 	btrfs_end_transaction(trans, root);
@@ -1237,7 +1240,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	mutex_lock(&root->fs_info->trans_mutex);
 	if (cur_trans->in_commit) {
-		cur_trans->use_count++;
+		atomic_inc(&cur_trans->use_count);
 		mutex_unlock(&root->fs_info->trans_mutex);
 		btrfs_end_transaction(trans, root);
 
@@ -1259,7 +1262,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		prev_trans = list_entry(cur_trans->list.prev,
 					struct btrfs_transaction, list);
 		if (!prev_trans->commit_done) {
-			prev_trans->use_count++;
+			atomic_inc(&prev_trans->use_count);
 			mutex_unlock(&root->fs_info->trans_mutex);
 
 			wait_for_commit(root, prev_trans);
@@ -1300,14 +1303,14 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 				TASK_UNINTERRUPTIBLE);
 
 		smp_mb();
-		if (cur_trans->num_writers > 1)
+		if (atomic_read(&cur_trans->num_writers) > 1)
 			schedule_timeout(MAX_SCHEDULE_TIMEOUT);
 		else if (should_grow)
 			schedule_timeout(1);
 
 		mutex_lock(&root->fs_info->trans_mutex);
 		finish_wait(&cur_trans->writer_wait, &wait);
-	} while (cur_trans->num_writers > 1 ||
+	} while (atomic_read(&cur_trans->num_writers) > 1 ||
 		 (should_grow && cur_trans->num_joined != joined));
 
 	ret = create_pending_snapshots(trans, root->fs_info);
@@ -1394,6 +1397,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	wake_up(&cur_trans->commit_wait);
 
+	list_del_init(&cur_trans->list);
 	put_transaction(cur_trans);
 	put_transaction(cur_trans);
 

fs/btrfs/transaction.h

@@ -27,11 +27,11 @@ struct btrfs_transaction {
 	 * total writers in this transaction, it must be zero before the
 	 * transaction can end
 	 */
-	unsigned long num_writers;
+	atomic_t num_writers;
 
 	unsigned long num_joined;
 	int in_commit;
-	int use_count;
+	atomic_t use_count;
 	int commit_done;
 	int blocked;
 	struct list_head list;

fs/btrfs/xattr.c

@@ -180,11 +180,10 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 	struct btrfs_path *path;
 	struct extent_buffer *leaf;
 	struct btrfs_dir_item *di;
-	int ret = 0, slot, advance;
+	int ret = 0, slot;
 	size_t total_size = 0, size_left = size;
 	unsigned long name_ptr;
 	size_t name_len;
-	u32 nritems;
 
 	/*
 	 * ok we want all objects associated with this id.
@@ -204,34 +203,24 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
-	advance = 0;
+
 	while (1) {
 		leaf = path->nodes[0];
-		nritems = btrfs_header_nritems(leaf);
 		slot = path->slots[0];
 
 		/* this is where we start walking through the path */
-		if (advance || slot >= nritems) {
+		if (slot >= btrfs_header_nritems(leaf)) {
 			/*
 			 * if we've reached the last slot in this leaf we need
 			 * to go to the next leaf and reset everything
 			 */
-			if (slot >= nritems-1) {
-				ret = btrfs_next_leaf(root, path);
-				if (ret)
-					break;
-				leaf = path->nodes[0];
-				nritems = btrfs_header_nritems(leaf);
-				slot = path->slots[0];
-			} else {
-				/*
-				 * just walking through the slots on this leaf
-				 */
-				slot++;
-				path->slots[0]++;
-			}
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto err;
+			else if (ret > 0)
+				break;
+			continue;
 		}
-		advance = 1;
 
 		btrfs_item_key_to_cpu(leaf, &found_key, slot);
 
@@ -250,7 +239,7 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 
 		/* we are just looking for how big our buffer needs to be */
 		if (!size)
-			continue;
+			goto next;
 
 		if (!buffer || (name_len + 1) > size_left) {
 			ret = -ERANGE;
@@ -263,6 +252,8 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 
 		size_left -= name_len + 1;
 		buffer += name_len + 1;
+next:
+		path->slots[0]++;
 	}
 	ret = total_size;